Three K2 Campaigns Yield Rotation Periods for 1013 Stars in Praesepe

We use three campaigns of K2 observations to complete the census of rotation in low-mass members of the benchmark, $\approx$670-Myr-old open cluster Praesepe. We measure new rotation periods (\prot) for 220 $\lesssim$1.3~\Msun\ Praesepe members and recover periods for $97\%$ (793/812) of the stars with a \prot\ in the literature. Of the 19 stars for which we do not recover a \prot, 17 were not observed by K2. As K2's three Praesepe campaigns took place over the course of three years, we test the stability of our measured \prot\ for stars observed in more than one campaign. We measure \prot\ consistent to within $10\%$ for $>95\%$ of the 331 likely single stars with $\geq$2 high-quality observations; the median difference in \prot\ is $0.3\%$, with a standard deviation of $2\%$. Nearly all of the exceptions are stars with discrepant \prot\ measurements in Campaign 18, K2's last, which was significantly shorter than the earlier two ($\approx$50~d rather than $\approx$75~d). This suggests that, despite the evident morphological evolution we observe in the light curves of $38\%$ of the stars, \prot\ measurements for low-mass stars in Praesepe are stable on timescales of several years. A \prot\ can therefore be taken to be representative even if measured only once.